A Reference Linkage Map for Eucalyptus

Hudson et al. BMC Genomics 2012, 13:240 http://www.biomedcentral.com/1471-2164/13/240

RESEARCH ARTICLE Open Access A reference linkage map for Eucalyptus Corey J Hudson1*, Jules S Freeman1,2, Anand RK Kullan3, César D Petroli4, Carolina P Sansaloni4, Andrzej Kilian5, Frank Detering5, Dario Grattapaglia6, Brad M Potts1, Alexander A Myburg3 and René E Vaillancourt1

Abstract Background: Genetic linkage maps are invaluable resources in plant research. They provide a key tool for many genetic applications including: mapping quantitative trait loci (QTL); comparative mapping; identifying unlinked (i.e. independent) DNA markers for fingerprinting, population genetics and phylogenetics; assisting genome sequence assembly; relating physical and recombination distances along the genome and map-based cloning of genes. Eucalypts are the dominant tree species in most Australian ecosystems and of economic importance globally as plantation trees. The genome sequence of E. grandis has recently been released providing unprecedented opportunities for genetic and genomic research in the genus. A robust reference linkage map containing sequence-based molecular markers is needed to capitalise on this resource. Several high density linkage maps have recently been constructed for the main commercial forestry species in the genus (E. grandis, E. urophylla and E. globulus) using sequenced Diversity Arrays Technology (DArT) and microsatellite markers. To provide a single reference linkage map for eucalypts a composite map was produced through the integration of data from seven independent mapping experiments (1950 individuals) using a marker-merging method. Results: The composite map totalled 1107 cM and contained 4101 markers; comprising 3880 DArT, 213 microsatellite and eight candidate genes. Eighty-one DArT markers were mapped to two or more linkage groups, resulting in the 4101 markers being mapped to 4191 map positions. Approximately 13% of DArT markers mapped to identical map positions, thus the composite map contained 3634 unique loci at an average interval of 0.31 cM. Conclusion: The composite map represents the most saturated linkage map yet produced in Eucalyptus. As the majority of DArT markers contained on the map have been sequenced, the map provides a direct link to the E. grandis genome sequence and will serve as an important reference for progressing eucalypt research.

Background their value to genetics research has led to numerous link- Genetic linkage maps are valuable resources which can be age mapping projects being undertaken in plants. Detailed used to provide a framework for many genomic analyses. linkage maps have been produced for all of the world’s Linkage maps can be used to investigate the organisation staple cereal species [12], and in forest trees, linkage maps and evolution of genomes through comparative mapping have been produced for many of the most widely-planted [1-3] and serve as a basis for investigating phenotypic species due to their commercial importance as wood and traits of ecological and economic importance through the fibre crops [1,13,14]. localisation of quantitative trait loci [QTL; 4-6]. Subse- Grattapaglia and Sederoff [15] published the first gen- quently, QTL results may be used to help guide the selec- etic linkage map in the forest tree genus Eucalyptus in tion of candidate genes for association studies or be 1994. Subsequently, many mapping pedigrees have been applied in marker-assisted breeding programmes [7,8]. established for the purpose of linkage map construction Linkage maps can also be used to anchor physical maps and associated QTL analyses. More than 20 eucalypt and assist in the assembly of genome sequences [9-11]. genetic linkage maps have been reported with most The wide application of linkage maps in combination with being produced in the main commercially grown species, or their hybrids, from the Eucalyptus subgenus Symphyo- myrtus. Thus, the majority of linkage mapping projects * Correspondence: [email protected] 1School of Plant Science and CRC for Forestry, University of Tasmania, Private have focussed on E. grandis, E. urophylla and E. globulus Bag 55 Hobart, Tasmania 7001, Australia [reviewed in 16], while a smaller number of maps have Full list of author information is available at the end of the article

© 2012 Hudson et al., licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Hudson et al. BMC Genomics 2012, 13:240 Page 2 of 11 http://www.biomedcentral.com/1471-2164/13/240

also been produced for E. nitens [17], E. teriticornis effects; 38] and are more useful in some genetic ana- [18,19], E. camaldulensis [20] and for species in the lyses, such as estimating population genetic parameters closely related genus Corymbia [21]. (e.g. inbreeding levels), relative to dominant marker Many early eucalypt linkage maps were constructed types such as DArT. In addition, the DArT marker assay using random amplification of polymorphic DNA can be subject to cross-hybridization from duplicated (RAPD) and amplified fragment length polymorphism loci in the genome, although most such artifacts can be (AFLP) molecular markers [16,22]. However, the an- excluded by preselecting markers exhibiting Mendelian onymous nature of these dominant markers has limited segregation ratios in mapping pedigrees. the transfer of linkage information between studies At present, DArT markers have been used to construct [16,23]. More informative, codominant markers such as linkage maps in seven independent E. globulus and/or isozyme and random fragment length polymorphism E. grandis × E. urophylla hybrid family mapping pedi- (RFLPs) have also been used in eucalypt linkage map- grees [11,32,33]. All of these maps also contain a variable ping, although, their low throughput, low inter-pedigree number of co-dominant microsatellite markers, which polymorphism and labour intensive genotyping require- provide important links to many earlier eucalypt linkage ments have limited their use [16,23]. The more recent maps. In the two largest mapping pedigrees (more than development of highly polymorphic microsatellite mar- 500 individuals each), 1010 [32] and 2229 [33] DArT kers made available a large potential suite of markers markers, were mapped at sub-centiMorgan marker that are transferrable between species and polymorphic densities and collectively more than 4000 DArT and in multiple pedigrees. This enabled linkage group syn- microsatellite markers have been mapped in the seven teny to be established between maps containing com- pedigrees. mon microsatellite markers and the positions and All DArT marker based linkage maps were constructed stability of QTL across multiple species to be examined using the program JoinMap 4.0 [37]. This program is one [e.g. 24-27]. The ability to establish linkage group syn- of the most commonly used linkage mapping programs teny has also enabled moderate-density comparative and appears to be the only software available for building mapping studies [23,28]. linkage maps using the combined segregation data from Recent advances in molecular methods have led to multiple populations [39-41]. However, it is presently not high-throughput genotyping systems being developed feasible to combine the segregation data contained within [e.g. 29,30]. These have made it possible to quickly gen- the seven eucalypt mapping families describe above (col- erate many hundreds of markers in single mapping pedi- lectively 1950 individuals), and successfully order such grees and have helped facilitate the construction of high large numbers of markers within linkage groups (up density linkage maps [12]. Most recently in Eucalyptus, to ~ 500) due to computational limitations (Van Ooijen Diversity Arrays Technology [DArT; 31] has been used pers comm.). To circumvent the limitations of traditional to generate large numbers of molecular markers for gen- segregation-based methods of linkage map construction, etic linkage mapping in several mapping pedigrees alternative marker-merging strategies have been devel- [e.g. 11,32,33]. The eucalypt DArT markers are highly oped. A so-called ‘composite map’ can be produced in transferable across species from subgenus Symphyomyr- which markers from individual component maps are tus [34] and the high-throughput array-based genotyping merged into a single map based on their position relative system provides wide genome coverage [35]. A key bene- to common anchor loci. For example, the ‘neighbours’ fit of the Eucalyptus DArT markers is the public avail- marker-merging approach of Cone et al. [42] and the ability of the sequences of most of the 7680 markers marker-merging method implemented in the PhenoMap contained on the genotyping array [GenBank accession program (GeneFlow Inc. USA) have been used to success- numbers HR865291 - HR872186], thus making it pos- fully construct high density composite maps containing sible to anchor DArT markers directly to the reference several thousand markers in a number of plant species; in- E. grandis genome sequence [v1.0 released January 2011; cluding Sorghum [43], barley [41,44,45] and maize [42,46]. 36]. However, while the DArT technology offers many In this study, a marker-merging method was used to advantages, the DArT markers do suffer some limita- construct a high-density DArT and microsatellite marker tions due to their dominant nature. For example, the in- composite linkage map from seven independently con- complete segregation information provided by those structed maps. Recent comparative mapping analyses DArT markers segregating in a 3:1 ratio (intercross) using 236 to 393 markers shared between three of the results in an exponential increase of marker-ordering maps [see 32] showed that these linkage maps exhibited calculations compared to fully-informative co-dominant high synteny (> 93.4% markers occurring on the same markers [37]. Co-dominant markers also provide more linkage groups) and high colinearity (> 93.7% markers complete information in QTL mapping studies [e.g. having the same order within linkage groups). This indi- allowing estimation of additive and dominant allelic cated that it would be possible to merge markers from Hudson et al. BMC Genomics 2012, 13:240 Page 3 of 11 http://www.biomedcentral.com/1471-2164/13/240

several component maps into a single high quality map maps comprised 11 linkage groups in accordance with featuring robust marker-order together with very high the haploid chromosome number of Eucalyptus [48]. marker density. It is expected that this composite map Before building the composite map, marker names were will facilitate marker and map information exchange and standardised across maps, homologous linkage groups were serve as a valuable reference for species in the subgenus identified using common (anchor) loci and marker co- Symphyomyrtus. linearity between component maps was visually inspected in MapChart [49]. Map data was supplied for both frame- Methods work (1032-marker) and comprehensive (2484-marker) The following terms are used to describe the various maps built in the GU-Emb family [see 11]. Based on the types of linkage maps reported in this paper; (1) sex- level of marker-order agreement between linkage groups averaged map – a consensus of individually constructed from these maps with other component maps, either GU- male and female maps, built in a single family using seg- Emb framework (LG’s1,3,5,7and9)orcomprehensive regation data from both parents, (2) consensus map – a (LG’s2,4,6,8,10and11)linkagegroupswereincludedin consensus of multiple individually constructed male and composite map construction. Five linkage groups from female maps, built in multiple families (e.g. F2 double- three of the smaller E. globulus mapping families (Table 1) pseudo backcross) using segregation data from all of the were found to have substantial regions of non-colinearity families, and (3) composite map – an integrated map of (discordant marker-orders) with other component maps. multiple sex-averaged and/or consensus maps, built Consequently, LG6 and LG10 from the GLOB-F1-1 map, using a marker-merging method. LG4 and LG9 from the GLOB-F1-4 and LG4 from the GLOB-F1-5mapwereexcludedfromcompositemap Component maps construction. The composite map was built using an E. grandis × The number of markers included for composite map E. urophylla F2 double pseudo-backcross pedigree consen- construction ranged from 498 (GLOB-F1-4) to 2290 sus linkage map [both species from section Latoangulatae; (GU-SA; Table 1). In total, this consisted of 4350 indi- 33] plus one E. grandis × E. urophylla sex-averaged map vidual markers, including: 4089 DArT, 253 microsatel- constructed in a F1 hybrid pedigree [11] and five pure- lites and eight mapped genes. Ninety-six markers (2.2% species E. globulus [section Maidenaria; 32] sex-averaged of the total number of markers; termed ‘multicopy’ mar- linkage maps constructed in either outcrossed F2 or F1 fam- kers) were mapped to two or more linkage groups across ilies (hereafter referred to as ‘component’ maps). Compo- component maps. This resulted in the 4350 individual nent map family sizes ranged from 172 (GLOB-F2-1) to markers being mapped to 4457 positions. Of these 4457 547 (GU-SA) and collectively contained 1,950 indivi- positions, 1960 could be considered to be bridging loci, duals (Table 1). The component maps were constructed meaning that these markers had been mapped to syn- by different researchers. All used JoinMap 4.0 [37] with tenic linkage groups in two or more component maps marker-ordering within linkage groups (LGs) estimated and would serve as anchor loci during composite map using the regression algorithm of Stam [47] combined construction. Conversely, 2497 marker positions were with the Kosambi mapping function. All component unique to single component maps.

Table 1 Component map details Linkage mapa Map n cM MMI Markers mapped abbreviation (percentage of unique markers in pedigree) DArT SSR Gene Total b E. grandis × E. urophylla SA double pseudo-backcross F2 GU-SA 547 1107 0.51 2229 (45%) 59 (46%) 2 (100%) 2290 (45%) ce E. grandis × E. urophylla Embrapa F1 GU-Emb 177 1229 0.78 1617 (41%) 193 (77%) 0 1810 (44%) d E. globulus Lighthouse F2 GLOB-LH 503 1151 1.21 1010 (27%) 50 (12%) 0 1060 (27%) d E. globulus FAM1 F1 GLOB-F1-1 184 1033 1.97 571 (14%) 4 (0%) 2 (0%) 577 (14%) d E. globulus FAM4 F1 GLOB-F1-4 184 1137 2.46 488 (10%) 6 (0%) 4 (25%) 498 (10%) d E. globulus FAM5 F1 GLOB-F1-5 183 1055 2.09 600 (22%) 4 (0%) 2 (0%) 606 (21%) d E. globulus FAM1 F2 GLOB-F2-1 172 1258 2.73 660 (18%) 30 (30%) 5 (40%) 695 (18%) Summary of the component maps used to construct the composite map. For each map, progeny size (n), map length (cM; total for all 11 linkage groups), mean marker interval (MMI; average for all 11 linkage groups) and total number of mapped markers (using only those linkage groups included in composite map construction; see Methods) are given. For DArT, microsatellite (SSR) and gene markers mapped on each component map, the percentage of markers unique to that map (i.e. not mapped in any of the six other component maps) are given in parentheses. aCross details and reference; bKullan et al. [33], cPetroli et al. [11] d e and Hudson et al. [32]. Data for the E. grandis × E. urophylla Embrapa F2 component map calculated using a combination of framework and comprehensive linkage groups (see Methods). Hudson et al. BMC Genomics 2012, 13:240 Page 4 of 11 http://www.biomedcentral.com/1471-2164/13/240

Composite map construction Composite map features The composite linkage map was constructed at Diver- Following composite map construction, marker-order sity Arrays Technology (DArT) Pty Ltd (Canberra, correlations between composite and component map Australia) using specially developed R scripts which linkage groups were calculated in SAS 9.2 (SAS Institute, merged component map markers into the composite Cary, USA) using the PROC CORR Spearman function. map based on their relative map positions. The E. To test whether multicopy markers were distributed 2 grandis × E. urophylla SA F2 (GU-SA) linkage map equally across linkage groups, a χ test was used to com- was used as the seed-map (i.e. the ‘fixed backbone’ to pare the observed versus expected number of multicopy which markers from other component maps were marker positions occurring on each linkage group. The added) due to it having the largest progeny size, the expected number of multicopy markers per linkage largest number of both mapped and unique markers group was calculated as; (total number of multicopy (Table 1) and high overall marker colinearity to the marker positions in the composite map/total number of 11 main superscaffolds of the assembled E. grandis DArT marker positions in the composite map) × number genome sequence [33,36]. The procedure for building of DArT marker positions per linkage group for that each composite map linkage group was as follows. linkage group. The BLAST server available at Phytozome Firstly, the number of common markers in each seed- [36] was used to search for DArT marker duplications. map – component map linkage group comparison The bl2seq tool at NCBI [50] was used to examine was identified. Spearman rank marker-order correla- DArT marker sequence similarity/redundancy. All tions were then estimated and a heuristic ‘fit value’ graphical representations of linkage maps were drawn for each comparison was calculated as; Fit value = cor- using MapChart [49]. relation×log (number of common markers); where the second term rewards for the number of common Results markers with a diminishing returns function. Follow- Composite map details ing selection of the component map linkage group A total of 4101 individual markers, comprising 3880 with the highest Fit value, unique markers (i.e. those DArT markers, eight gene-based markers and 213 not mapped on the seed linkage group, or the ‘build- microsatellite markers were included in the composite ing’ composite linkage group in following rounds) map. The composite map totalled 1107 cM which was were added to the seed linkage group (or ‘building’ within the range of component map lengths (1033– composite map linkage group) using linear regression. 1258 cM; Table 1) and contained only eleven marker Here, the slope (m) and intersect (c) calculated from intervals ≥ 3 cM; with a maximum marker interval of fitting the positions of common markers on the seed 5.9 cM. The composite map contained 81 multicopy linkage group (pc) to their positions on the selected DArT markers (2.1% of total DArT markers) which were component map (pi) linkage group (pc = m × pi + c) mapped to 171 map positions. Most multicopy markers was used to calculate the positions of unique compo- occurred on two linkage groups only, however, one mar- nent map markers added to the seed linkage group. ker (ePt-574238) mapped to three linkage groups while Once this first round was completed, the remaining four markers (ePt-503174, ePt-568818, ePt-637610, ePt- component linkage groups were compared to this 637861) mapped to four linkage groups. This resulted in new ‘building’ composite map linkage group and the the 4101 markers being mapped to 4191 positions process was repeated. This continued until all unique (Table 2). Over half (2171 or 53%) of the markers markers had been added from remaining component mapped to these 4191 map positions had been mapped maps which shared at least three common markers in a single component map only (i.e. were not shared with the building composite map linkage group and among multiple component maps). Approximately 13% had a marker-order correlation coefficient ≥ 0.50. This of DArT markers mapped to identical positions in the process was repeated for each linkage group to yield composite map. Therefore, the map contained 3634 the final composite map of 11 linkage groups. Mar- unique map loci with an average interval of 0.31 cM. kers which mapped to the distal ends of composite The number of multicopy DArT marker positions on linkage groups and which had relatively large inter- each linkage group ranged from 5 to 24 and represented marker intervals (≥ 5 cM) and poor support (e.g. 1.9-6.4% of the total number of DArT markers mapped mapped in one component map only) were removed. per linkage group (Table 2). Although LG5 and LG7 The numbering and orientation of linkage groups fol- contained a larger proportion of multicopy DArT mar- lowed the convention established in Brondani et al. ker positions (e.g. LG1 contained only 5 multicopy [23]; this also corresponds to the numbering of pseu- DArT marker positions, or 1.9% of the total number of dochromosome assemblies in the E. grandis genome DArT marker positions; Table 3), the proportion of mul- sequence [36]. ticopy DArT marker positions found on each linkage Hudson et al. BMC Genomics 2012, 13:240 Page 5 of 11 http://www.biomedcentral.com/1471-2164/13/240

Table 2 Composite map summary origins and multicopy DArT marker information is pre- LG cM Markers mapped Average MC sented in Additional file 1. marker DArT DArT SSR Genes Total a interval pos. – (cM) Composite component map colinearity Colinearity between component and composite map 1 93.8 250 12 0 262 0.42 5 linkage groups can be viewed graphically in Figure 1 2 102.1 451 29 0 480 0.24 18 (for the GLOB-LH map) and in Additional file 2 (all 3 105.6 429 18 2 449 0.28 21 component maps). Pair-wise linkage group marker- 4 80.9 219 9 3 231 0.41 12 order correlations were generally high (greater than 5 95.9 366 8 0 374 0.30 24 0.90; Table 3) reflecting the high colinearity shown 6 125.3 408 43 1 452 0.31 15 between common markers (Figure 1 and Additional file 2). However, a small degree of non-colinearity 7 87.7 305 9 1 315 0.33 18 did occur between all component maps and the com- 8 137.3 540 26 0 566 0.28 19 posite map. Eleven component map linkage groups 9 82.9 312 20 0 332 0.29 10 had marker-order correlations of less than 0.90 10 97.8 336 20 1 357 0.30 12 (Table 3), however, these linkage groups were either, 11 97.3 354 19 0 373 0.31 17 (1) identified as having poor marker colinearity with Total 1106.5 3970 213 8 4191 0.31 171b other component maps prior to composite map con- Summary of the composite map: including the number of mapped markers, struction and excluded from analysis (five linkage length and average marker intervals by linkage group (LG). aMC DArT pos. - groups with gray shading in Table 3), or (2) marker- indicates the number of multicopy (MC) DArT marker positions (pos.) occurring order information from these linkage groups was not on each linkage group. bThe 171 multicopy DArT marker positions represent 81 multicopy DArT markers (see Additional file 1). incorporated during composite map construction (correlation value without asterisk; six linkage groups group did not significantly differ from that expected by Table 3) due to markers from these maps being pre- chance across all linkage groups (χ2 = 12.99, P = 0.22, viously added from other linkage groups having bet- df = 10). There was no trend within linkage groups for ter fit values. Thus, these poorly correlated linkage multicopy DArT markers to be clumped in either distal groups did not adversely affect the composite map or central linkage group areas (data not shown). Com- marker-order. For each linkage group, the average posite map marker details, component map(s) marker pair-wise marker-order correlation between the

Table 3 Composite – component map marker-order correlation coefficients Composite map LG Component map a b GLOB-LH GU-Emb GLOB-F2-1 GLOB-F1-1 GLOB-F1-4 GLOB-F1-5 Average Average 1 0.98* 0.56F 0.99* 0.99* 0.98* 0.95* 0.91 0.98 2 0.98* 0.95C* 0.93* 0.95* 0.98* 0.85 0.94 0.96 3 0.91* 0.99F* 0.97* 0.99* 0.97* 0.99* 0.97 0.97 4 0.98* 0.74C 0.96* 0.89 0.79ex 0.19ns,ex 0.76 0.97 5 0.99* 0.92F 0.96* 0.96* 0.99* 0.96* 0.96 0.97 6 0.99* 0.99C* 0.99* 0.63ex 0.99* 0.86 0.91 0.99 7 0.98* 0.65F 0.99* 0.98* 0.96* 0.91* 0.91 0.96 8 0.98* 0.99C* 0.99* 0.66 0.94* 0.99* 0.93 0.98 9 0.99* 0.97F* 0.98* 0.97* 0.65ex 0.97 0.92 0.98 10 0.95* 0.98C* 0.97* 0.35ns,ex 0.92 0.97* 0.86 0.97 11 0.99* 0.99C* 0.99* 0.97* 0.96 0.99* 0.98 0.99 Averagec 0.98 0.88 0.97 0.85 0.92 0.87 Averaged 0.98 0.98 0.97 0.97 0.97 0.96 Marker-order correlations between composite map and component map linkage groups; the GU-SA component map is not shown as this map was used as the seed-map (i.e. provided a fixed-order and all correlations were 1.0). For the GU-Emb component map, superscript letters indicates whether the framework (F) or comprehensive (C) linkage group was used in map construction. Component map linkage groups initially excluded from composite map construction are indicated by ex superscript. An asterisk following the correlation value indicates that marker-order information from the component map was incorporated during construction of the composite map linkage group. Apart from two correlations (indicated by ns superscript) all correlations were significant at α ≤ 0.05. Averages: acalculated using all six component maps, bcalculated using only those linkage groups included in composite map construction (marked with an asterisk), ccalculated using all 11 linkage groups, dcalculated using only those linkage groups included in composite map construction (marked with an asterisk). Hudson et al. BMC Genomics 2012, 13:240 Page 6 of 11 http://www.biomedcentral.com/1471-2164/13/240

134562 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 78910 11 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Figure 1 Marker colinearity between the GLOB-LH component map (left) and composite map (right). Lines between each homologous linkage group pair indicate the positions of common markers. The scale bar is in Kosambi’s centiMorgans. composite map and those component maps included 606 base-pair similarity, e-value: 0.0). For the four in map construction ranged from 0.96 to 0.99 (Aver- unique multicopy markers, three were detected to have ageb column; Table 3). loci duplications within the E. grandis genome sequence. In each case, the positions of duplicated loci detected in DArT marker duplications the E. grandis genome sequence corresponded to the Although not a main focus of this study, evidence for linkage groups to which the marker was mapped. the occurrence of duplicated DArT marker loci within the assembled E. grandis genome sequence [36] was Discussion investigated for the five multicopy markers which had Composite map construction been mapped to three or more linkage groups. Two of Data from seven component maps were integrated into these markers (ePt-637610 and ePt-637861; see Add- a single composite map which represents the highest itional file 1) mapped to the same map position on each density map yet produced in Eucalyptus. A major advan- of four linkage groups (LGs 2, 3, 5 and 8) and were tage of the marker-merger method used in this study found to be redundant markers (i.e. identical sequences) was the substantial time and labour savings made when based on their marker sequence similarity (bl2seq: 583/ compared to the effort required to produce comparable Hudson et al. BMC Genomics 2012, 13:240 Page 7 of 11 http://www.biomedcentral.com/1471-2164/13/240

maps using traditional, segregation-based methods. For which are evolutionary meaningful units compared to base example, Li et al. [40] constructed a 2111 marker com- pair distances. Additionally, it is also important to under- posite map from four barley mapping pedigrees and stand the relationship between physical map (i.e. genome) reported that it took ‘several thousand hours’ of comput- and genetic map distances as this can have implications for ing time. In a larger barley study, Wenzl et al. [41] pro- map-based cloning efforts and/or marker-assisted selection. duced a 2935 loci composite map from ten mapping For example, uneven recombination rates across a genome populations using JoinMap 3.0 [51] in combination with [12,55] may result in physically distant markers appearing specially built Perl scripts and reported that the project to be genetically close to each other, or vice versa. In euca- required several months of semi-manual data processing lypts, Kullan et al. [33] recently compared 153 linkage map [41]. In contrast, the composite map produced in this intervals of approximately 1 cM against contigs of the study was built in a single day. E. grandis genome and found that the genetic map to physical distance relationship varied considerably; ranging from Utility of the composite map 100 kb to 2.4 Mbp per 1 cM. Therefore, the composite As sequences are available for the majority of DArT mar- map will be useful to provide further insight into the rela- kers on the map (91%; data not shown), the composite tionship between physical and genetic map distance in map provides a direct link to the E. grandis genome se- addition to identifying hot (or cold) spots of recombination. quence [36]. We have made use of this link to search the A key use of the composite map will be for compari- E. grandis genome sequence for candidate genes asso- son of QTL and candidate gene positions detected ciated with QTL locations and to facilitate the placement across variable genetic backgrounds and/or environ- of candidate genes in the component linkage maps with- ments in different studies. This has previously been lim- out the need for time consuming marker development ited due to a lack of common markers being shared and genotyping. Sequence-based linkage maps have also between maps [23]. For example, Thumma et al. [27] provided useful tools to aid in the assembly of genome detected multiple co-locating growth-related QTL on sequences [e.g. 52,53] and can be particularly beneficial in LG5 in E. nitens but could not accurately compare the taxa (such as eucalypts) which have a relatively small gen- position of this QTL to similar growth-related trait QTL ome size. For example, during the assembly of the detected on this same linkage group in two other studies E. grandis genome sequence, a DArT linkage map was [24,56]. Although most of the markers contained on the valuable in guiding contigs into the 11 main pseudochro- composite map are DArT markers, which to date have mosomes [16]. However, not all contigs could be aligned only been mapped in the pedigrees included in this and approximately 12% of the 693 Mbp E. grandis genome study, the map does contain several hundred microsatel- sequence remains unassembled in more than 4900 small lite markers (213) which will enable synteny and co- unlinked scaffolds [54]. With the composite map contain- linearity to be established with many earlier linkage ing many more DArT markers (1600+) than the linkage maps used for QTL detection; e.g. 13 out of 22 earlier map used to aid genome assembly, the composite map studies have mapped a variable number of microsatel- markers may provide further positional information and lites [16]. This will enable QTL to be aligned against the help to anchor some of the unlinked scaffolds and refine composite map which may provide deeper insight into the current E. grandis genome sequence. the genetic control of phenotypic traits in the genus. For Over half (53%) of the markers placed in the composite example, following the construction of an integrated map originated from a single component map (i.e. were map for melon (Cucumis melo) which used data from not shared among multiple component maps). Therefore, eight independent mapping experiments, it was possible the ability to determine the relative positions of markers to align 370 QTL detected for 62 traits from 18 experi- mapped in different maps has been greatly enhanced ments [57]. Through this alignment, QTL detected in through the integration of this data into a single map. This different studies for economically important traits were has already proven advantageous to our research group, found to co-locate [57]; providing supporting evidence with the composite map being used to quickly identify the to substantiate the biological basis of the observed linkage relationships of microsatellite markers used in marker-trait association [7,8]. population genetic studies. Although now a relatively sim- As in all linkage mapping studies, it is important to con- ple task, it was previously necessary to consult multiple sider both the quality of the map produced and any spe- linkage maps and assess their colinearity to obtain this cific map characteristics. In the alignment of 6480 DArT same information. Furthermore, any marker developed in marker sequences against the E. grandis genome sequence eucalypts which has known sequence, can now potentially [36], Petroli et al. [11] reported that although the majority be found in the eucalypt genome sequence and then of markers (4189) occurred at a single genome position aligned against the reference map in order to estimate its with high support, many marker sequences (2291), albeit distance to other markers in units of recombination (cM); at lower confidence, also exhibited similarity to a second Hudson et al. BMC Genomics 2012, 13:240 Page 8 of 11 http://www.biomedcentral.com/1471-2164/13/240

genome position and that about half of these genome high correlations, most component maps did exhibit some regions contained repeat elements. Furthermore, prelim- marker-order inconsistencies with the composite map. A inary analysis of the E. grandis genome sequence suggests number of (mostly) single marker-order inconsistencies that (as has been observed in some Rosid genomes) a did occur over large distances, but most marker-order dis- whole-genome duplication event has occurred in the agreements occurred among tightly grouped markers in lineage (Myrtales) subsequent to the ancient hexaploidy regions of less than 5 cM. Although it is possible that event shared by all rosids (Myburg et al., unpublished). some of these marker-order differences could be real and Such whole-genome, as well as, segmental duplication represent local chromosomal rearrangements or marker events will affect thousands of marker loci, but most duplications between the different mapping pedigrees would be expected to diverge in sequence with evolution- and/or species, they are more likely to reflect marker- ary time yielding mostly unique marker loci. Thus, the order inaccuracies within any of the component maps or presence of multicopy markers (representing putatively simply be artefacts of the statistical uncertainty associated duplicated loci) in the composite map was not unex- with ordering tightly linked markers [see 58]. While users pected. It is worth noting that in the construction of each of this map should be aware of these limitations and how component map, only those markers which segregated as they may affect marker ordering, overall, the generally a single Mendelian locus were mapped. Therefore, in the high marker-order correlations observed and the exclu- event of a marker duplication being present within a pedi- sion of component map linkage groups having poor mar- gree, only one locus could be polymorphic in order for ker colinearity from initial composite map construction that marker to produce a single loci segregation ratios. (and thus not adversely affecting composite map marker- Consequently, it is likely that only a subset of the dupli- order) suggests that the composite map is of a sufficiently cated loci present within the eucalypt genome have been high quality to facilitate the transfer of genetic information identified in the composite map. Given that the PstI between studies. enzyme used in the complexity reduction step of DArT The composite map will be most useful for studies in- marker development [35] preferentially produces markers volving species from subgenus Symphyomyrtus sections located in hypomethylated, gene rich regions [55], and Latoangulatae and Maidenaria; due to the composite that many DArT markers contain protein coding map being built from linkage maps constructed in spe- sequences [33], it is possible that some of the multicopy cies from these sections. However, due to the high level markers identified may be associated with different gene of genome synteny and colinearity detected between family members and/or be part of larger duplicated species from these relatively distant sections [28,32,34], regions. Further studies are required to examine the full information from the composite map should also be ap- extent and evolution of the duplicated loci. We also plicable to many other commercially important eucalypt expected some marker redundancy (markers with the species in closely related sections (e.g. E. camaldulensis same sequence) among the 3808 composite map DArT from subgenus Symphyomyrtus section Exsertaria). markers; an issue which arises due to the process by which DArT markers are generated, resulting in the same ampli- Future marker integration fied genomic fragment being represented more than once A number of recent studies have focussed on the develop- on the genotyping array [31,35]. Therefore, identical ment of molecular markers for use in eucalypts. In clones (e.g. the same DArT fragment, but with different addition to the DArT genotyping array developed for use DArT marker names) are expected to produce identical in eucalypts [35], the feasibility of high-throughput SNP genotype scores and should map to identical (or near genotyping has been explored [59] and several tens of identical) map positions; as found for the markers ePt- species-transferrable EST-based SSR markers have been 637610 and ePt-637861 identified as identical clones in recently reported [60,61]. Furthermore, DArT genotyping this study. by sequencing (GBS), which combines the complexity The marker-merging method used in this study took ad- reduction method of DArT [31] with next generation se- vantage of the fact that individual component maps were quencing (NGS), and which can potentially deliver up to constructed using high marker-ordering stringency which three-fold as many markers as conventional DArT geno- resulted in linkage maps having robust marker-orders typing [see 62] is becoming a cost-competitive genotyping [32]. The comparison of the composite map marker-order option due to the recent plummeting costs of NGS se- against individual component maps gives an indication of quencing. Therefore, to broaden the use of the composite the quality of the composite map. Marker-order correla- map for comparative analyses and to optimise its’ worth, it tions were mostly excellent with high pair-wise linkage will be necessary to add new markers to the current ver- group marker-order correlations found in most compari- sion of the composite map in the future. Although beyond sons. For example, in 48 out of 66 pair-wise comparisons the scope of this study, it would also be valuable to com- the marker-order correlation exceeded 0.95. Despite these pare the marker order of the composite map to maps built Hudson et al. BMC Genomics 2012, 13:240 Page 9 of 11 http://www.biomedcentral.com/1471-2164/13/240

using the same data with other marker-merging software respectively. AK and FD constructed the composite map. REV and AAM (e.g. BioMercator [63], CarthaGene [64] or MergeMap conceived the study, and along with BMP and JSF, contributed to the design of the study. All authors read and approved the final manuscript. [65]). The R scripts and map marker positions of the component maps used in this study can be made available Funding upon request. Funding for this project was provided by the Australian Research Council (DP0770506 & DP110101621) as well as the Cooperative Research Centre for Conclusion Forestry (Australia). Construction of the GU-SA map was supported by Sappi, Mondi, the Technology and Human Resources for Industry Program (THRIP), The integration of markers from seven individual genetic the National Research Foundation (NRF) and the Department of Science and linkage pedigrees has resulted in a composite, reference Technology (DST) of South Africa. map for eucalypts with 4101 DArT and microsatellite Author details markers. Although some small marker-order inconsist- 1School of Plant Science and CRC for Forestry, University of Tasmania, Private encies exist between component maps and the compos- Bag 55 Hobart, Tasmania 7001, Australia. 2CRN Research Fellow, Faculty of ite map, there is a relatively high agreement of marker- Science, Health, Education and Engineering, University of the Sunshine Coast, Locked Bag 4, Maroochydore, QLD 4558, Australia. 3Department of Genetics, order between component maps; which indicates that Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, the composite map represents a good estimation of the Pretoria 0002, South Africa. 4EMBRAPA Genetic Resources and Biotechnology true marker positions in most cases. However, at finer - EPqB Final W5 Norte 70770–917 Brazilia DF and Dep. Cell Biology, Universidade de Brazilia – UnB, Brasilia, DF, Brazil. 5Diversity Arrays scales (sub-cM) marker-orders may differ between com- Technology Pty Ltd, PO Box 7141, Yarralumla, ACT 2600, Australia. 6EMBRAPA ponent and composite maps due to limited statistical Genetic Resources and Biotechnology - Parque Estação Biológica - PqEB - Av. power to order such tightly linked markers. Overall, the W5 Norte (final), Brasília, DF - Brazil - 70770–917, Universidade Catolica de Brasília- SGAN, 916 modulo B, 70790-160DF, Brasilia, Brazil. genome coverage and marker density of the composite map greatly exceeded that achieved in any of the single Received: 23 March 2012 Accepted: 4 June 2012 mapping populations. It is expected that this composite Published: 15 June 2012 map will provide a valuable reference map for the world-wide Eucalyptus research community, facilitate References 1. Krutovsky KV, Troggio M, Brown GR, Jermstad KD, Neale DB: Comparative the transfer of genetic information between different mapping in the Pinaceae. Genetics 2004, 168:447–461. studies and allow for the integration of DArT marker in- 2. Lefebvre-Pautigny F, Wu F, Philippot M, Rigoreau M, Zouine M, Frasse P, formation with other genomic resources. Bouzayen M, Broun P, Pétiard V, et al: High resolution synteny maps allowing direct comparisons between the coffee and tomato genomes. Tree Genetics & Genomes 2009, 6:565–577. Additional files 3. Paterson AH, Bowers JE, Burow MD, Draye X, Elsik CG, Jiang C-X, Katsar CS, Lan T-H, Lin Y-R, Ming R, Wright RJ: Comparative genomics of plant chromosomes. Plant Cell 2000, 12:1523–1540. Eucalyptus Additional file 1: composite map details. Details of 4. Anderson JT, Lee C-R, Mitchell-Olds T: Life-history QTLs and natural markers mapped in the Eucalyptus composite map. Includes, linkage selection on flowering time in Boechera stricta, a perennial relative of group and position of mapped markers, marker type and which Arabidopsis. Evolution 2011, 65:771–787. component map(s) markers were mapped. A '1' in the 'Multicopy marker' 5. Freeman JS, O'Reilly-Wapstra JM, Vaillancourt RE, Wiggins N, Potts BM: column indicates that the marker occurs on two or more linkage groups. Quantitative trait loci for key defensive compounds affecting herbivory Additional file 2: Composite map – component map marker of eucalypts in Australia. New Phytologist 2008, 178:846–851. colinearity. Marker colinearity among all six component maps and the 6. Kearsey MJ, Farquhar AGL: QTL analysis in plants; where are we now? Eucalyptus composite map. For each linkage group, three linkage group Heredity 1998, 80:137–142. ‘triplets’ show marker colinearity between two component maps 7. Brown GR, Bassoni DL, Gill GP, Fontana JR, Wheeler NC, Megraw RA, Davis MF, (outside) and the composite map (centre). Horizontal lines on linkage Sewell MM, Tuskan GA, Neale DB: Identification of quantitative trait loci group bars indicate marker positions and lines between linkage groups influencing wood property traits in loblolly pine (Pinus taeda L.). III. QTL indicate the position of common markers. The scale bar shown is in verification and candidate gene mapping. Genetics 2003, 164:1537–1546. Kosambi’s centiMorgans. Component map names (abbreviations; see 8. Wheeler NC, Jermstad KD, Krutovsky K, Aitken SN, Howe GT, Krakowski J, Table 1) are given above each linkage group. Linkage groups excluded Neale DB: Mapping of quantitative trait loci controlling adaptive traits in from composite map construction are indicated in parentheses following coastal Douglas-fir. IV. Cold-hardiness QTL verification and candidate the component map name. An asterisk indicates whether marker-order gene mapping. Molecular Breeding 2005, 15:145–156. information from the component map was incorporated during 9. Semagn K, Bjornstad A, Ndjiondjop MN: Principles, requirements and composite map construction (see Methods). For the GU-Emb component prospects of genetic mapping in plants. African Journal of Biotechnology map, superscript letters indicates whether the framework (f) or 2006, 5:2569–2587. comprehensive (c) linkage group from this pedigree was used in 10. The Potato Genome Sequencing Consortium: Genome sequence and composite map construction. analysis of the tuber crop potato. Nature 2011, 475:189–195. 11. Petroli C, Sansaloni C, Carling J, Mamani E, Steane D, Myburg A, Vaillancourt R, Kilian A, Pappas G, Bonfim da Silva O, Grattapaglia D: Genomic Competing interest characterization, high-density mapping and anchoring of DArT markers to 'The authors declare that they have no competing interests. the reference genome of Eucalyptus. BMC Proceedings 2011, 5(Suppl 7):P35. 12. Jones N, Ougham H, Thomas H, Pašakinskienė I: Markers and mapping Authors’ contributions revisited: finding your gene. New Phytologist 2009, 183:935–966. CJH built the GLOB-LH linkage map, coordinated the collection of map data, 13. Grattapaglia D, Plomion C, Kirst M, Sederoff RR: Genomics of growth traits building of the composite map, performed analyses and wrote the in forest trees. Current Opinion in Plant Biology 2009, 12:148–156. manuscript. JSF built all other E. globulus linkage maps. AAM and ARKK, DG 14. Poke F, Vaillancourt RE, Potts B, Reid J: Genomic research in Eucalyptus. and CDP, built and contributed GU-SA and GU-Emb linkage maps, Genetica 2005, 125:79–101. Hudson et al. BMC Genomics 2012, 13:240 Page 10 of 11 http://www.biomedcentral.com/1471-2164/13/240

15. Grattapaglia D, Sederoff R: Genetic linkage maps of Eucalyptus grandis 36. Eucalyptus grandis genome (JGI v1.0). http://www.phytozome.net/cgi-bin/ and Eucalyptus urophylla using a pseudo-testcross: mapping strategy gbrowse/eucalyptus/. and RAPD markers. Genetics 1994, 137:1121–1137. 37. Van Ooijen J: JoinMap 4, software for the calculation of genetic linkage maps 16. Grattapaglia D, Vaillancourt RE, Shepherd M, Thumma BR, Foley W, Kulheim in experimental populations. Wageningen, Netherlands: Kyazma B.V; 2006. C, Potts BM, Myburg AA: Progress in Myrtaceae genomics: Eucalyptus as 38. Grattapaglia D: Molecular breeding of Eucalyptus.InIn Molecular biology of the pivotal genus. Tree Genetics & Genomes, 8:463–508. in press. woody plants Volume 1. Edited by Jain S, Minocha S. Netherlands: Kluwer; 17. Byrne M, Murrell JC, Allen B, Moran GF: An integrated genetic linkage map 2000:451–474. for eucalypts using RFLP, RAPD and isozyme markers. Theoretical and 39. Cheema J, Dicks J: Computational approaches and software tools for Applied Genetics 1995, 91:869–875. genetic linkage map estimation in plants. Briefings in Bioinformatics 2009, 18. Gan S, Shi J, Li M, Wu K, Wu J, Bai J: Moderate-density molecular maps of 10:595–608. Eucalyptus urophylla (S. T. Blake) and E. tereticornis (Smith) genomes 40. Li H, Kilian A, Zhou M, Wenzl P, Huttner E, Mendham N, McIntyre L, based on RAPD markers. Genetica 2003, 118:59–67. Vaillancourt R: Construction of a high-density composite map and 19. Marques CM, Araújo JA, Ferreira JG, Whetten R, O’Malley DM, Liu BH, comparative mapping of segregation distortion regions (SDRs) in barley. Sederoff R: AFLP genetic maps of Eucalyptus globulus and E. tereticornis. Molecular Genetics and Genomics 2010, 284:319–331. Theoretical and Applied Genetics 1998, 96:727–737. 41. Wenzl P, Li H, Carling J, Zhou M, Raman H, Paul E, Hearnden P, Maier C, Xia 20. Agrama HA, Salah SF: Construction of a genome map for Eucalyptus L, Caig V, et al: A high-density consensus map of barley linking DArT camaldulensis DEHN. Silvae Genetica 2002, 51:201–206. markers to SSR. RFLP and STS loci and agricultural traits. BMC Genomics 21. Shepherd M, Kasem S, Lee D, Henry R: Construction of microsatellite 2006, 7:206. linkage maps for Corymbia. Silvae Genetica 2006, 55:228–238. 42. Cone KC, McMullen MD, Bi IV, Davis GL, Yim Y-S, Gardiner JM, Polacco ML, 22. Myburg AA, Potts BM, Marques CM, Kirst M, Gion J, Grattapaglia D, Sanchez-Villeda H, Fang Z, Schroeder SG, et al: Genetic, physical, and Grima-Pettenatti J: Eucalypts.InGenome mapping and molecular breeding in informatics resources for Maize. On the road to an integrated map. Plant plants. Volume 7. Edited by Kole C. Berlin: Springer; 2007:115–160. Physiology 2002, 130:1598–1605. 23. Brondani RPV, Williams ER, Brondani C, Grattapaglia D: A microsatellite-based 43. Mace E, Rami J-F, Bouchet S, Klein P, Klein R, Kilian A, Wenzl P, Xia L, consensus linkage map for species of Eucalyptus and a novel set of 230 Halloran K, Jordan D: A consensus genetic map of sorghum that microsatellite markers for the genus. BMC Plant Biology 2006, 6:20. integrates multiple component maps and high-throughput Diversity 24. Freeman J, Whittock S, Potts B, Vaillancourt R: QTL influencing growth and Array Technology (DArT) markers. BMC Plant Biology 2009, 9:13. wood properties in Eucalyptus globulus. Tree Genetics & Genomes 2009, 44. Alsop B, Farre A, Wenzl P, Wang J, Zhou M, Romagosa I, Kilian A, Steffenson B: 5:713–722. Development of wild barley-derived DArT markers and their integration 25. Gion J-M, Carouche A, Deweer S, Bedon F, Pichavant F, Charpentier J-P, into a barley consensus map. Molecular Breeding 2011, 27:77–92. Bailleres H, Rozenberg P, Carocha V, Ognouabi N, et al: Comprehensive 45. Varshney R, Marcel T, Ramsay L, Russell J, Röder M, Stein N, Waugh R, genetic dissection of wood properties in a widely-grown tropical tree: Langridge P, Niks R, Graner A: A high density barley microsatellite Eucalyptus. BMC Genomics 2011, 12:301. consensus map with 775 SSR loci. Theoretical and Applied Genetics 2007, 26. Marques C, Brondani R, Grattapaglia D, Sederoff R: Conservation and 114:1091–1103. synteny of SSR loci and QTLs for vegetative propagation in four 46. Wei F, Coe E, Nelson W, Bharti AK, Engler F, Butler E, Kim H, Goicoechea JL, Eucalyptus species. Theoretical and Applied Genetics 2002, Chen M, Lee S, et al: Physical and genetic structure of the Maize genome 105:474–478. reflects its complex evolutionary history. PLoS Genet 2007, 3:e123. 27. Thumma B, Baltunis B, Bell J, Emebiri L, Moran G, Southerton S: Quantitative 47. Stam P: Construction of integrated genetic linkage maps by means of a trait locus (QTL) analysis of growth and vegetative propagation traits in new computer package: JoinMap. Plant J 1993, 3:739–744. Eucalyptus nitens full-sib families. Tree Genetics & Genomes 2010, 48. Bachir O, Abdellah B: Chromosome numbers of the 59 species of 6:877–889. Eucalyptus L'Herit (Myrtaceae). Caryologia 2006, 59:207–212. 28. Myburg AA, Griffin AR, Sederoff RR, Whetten RW: Comparative genetic 49. Voorrips RE: MapChart: software for the graphical presentation of linkage linkage maps of Eucalyptus grandis, Eucalyptus globulus and their F1 maps and QTLs. Journal of Heredity 2002, 93:77–78. hybrid based on a double pseudo-backcross mapping approach. 50. NCBI: NCBI BLAST. http://blast.ncbi.nlm.nih.gov/. Theoretical and Applied Genetics 2003, 107:1028–1042. 51. Van Ooijen J, Voorrips RE: JoinMap 3.0, software for the calculation of 29. Appleby N, Edwards D, Batley J: New technologies for ultra-high genetic linkage maps.InPlant Research International. Wageningen, throughput genotyping in plants.InMethods in Molecular Biology, Plant Netherlands; 2001. genomics. Volume 153. Edited by Somers DJ, Langridge P, Gustafson JP. New 52. Wang Y, Sun S, Liu B, Wang H, Deng J, Liao Y, Wang Q, Cheng F, Wang X, York: Humana Press; 2009:19–39. Wu J: A sequence-based genetic linkage map as a reference for Brassica 30. Davey JW, Hohenlohe PA, Etter PD, Boone JQ, Catchen JM, Blaxter ML: rapa pseudochromosome assembly. BMC Genomics 2011, 12:239. Genome-wide genetic marker discovery and genotyping using 53. Hwang T-Y, Sayama T, Takahashi M, Takada Y, Nakamoto Y, Funatsuki H, next-generation sequencing. Nat Rev Genet 2011, 12:499–510. Hisano H, Sasamoto S, Sato S, Tabata S, et al: High-density integrated 31. Jaccoud D, Peng K, Feinstein D, Kilian A: Diversity arrays: a solid state linkage map based on SSR markers in soybean. DNA Research 2009, technology for sequence information independent genotyping. Nucleic 16:213–225. Acids Research 2001, 29(4):e25. 54. Eucagen: Early release of the E. grandis genome sequence. http://web.up.ac.za/ 32. Hudson CJ, Kullan ARK, Freeman JS, Faria D, Grattapaglia D, Kilian A, Myburg eucagen/. A, Potts BM, Vaillancourt RE: High synteny and colinearity among 55. van Os H, Andrzejewski S, Bakker E, Barrena I, Bryan GJ, Caromel B, Ghareeb Eucalyptus genomes revealed by high-density comparative genetic B, Isidore E, de Jong W, van Koert P, et al: Construction of a 10,000-marker mapping. Tree Genetics & Genomes 2012, 8:339–352. ultradense genetic recombination map of potato: providing a framework 33. Kullan A, van Dyk M, Jones N, Kanzler A, Bayley A, Myburg A: High-density for accelerated gene isolation and a genomewide physical map. Genetics genetic linkage maps with over 2,400 sequence-anchored DArT markers 2006, 173:1075–1087. for genetic dissection in an F2 pseudo-backcross of Eucalyptus grandis x 56. Grattapaglia D, Bertolucci FLG, Penchel R, Sederoff RR: Genetic mapping of E. urophylla. Tree Genetics & Genomes 2012, 8:163–175. quantitative trait loci controlling growth and wood quality traits in 34. Steane DA, Nicolle D, Sansaloni CP, Petroli CD, Carling J, Kilian A, Myburg AA, Eucalyptus grandis using a maternal half-sib family and RAPD markers. Grattapaglia D, Vaillancourt RE: Population genetic analysis and phylogeny Genetics 1996, 144:1205–1214. reconstruction in Eucalyptus (Myrtaceae) using high-throughput, 57. Diaz A, Fergany M, Formisano G, Ziarsolo P, Blanca J, Fei Z, Staub J, Zalapa J, genome-wide genotyping. Molecular Phylogenetics and Evolution 2011, Cuevas H, Dace G, et al: A consensus linkage map for molecular markers 59:206–224. and quantitative trait loci associated with economically important traits 35. Sansaloni C, Petroli C, Carling J, Hudson C, Steane D, Myburg A, Grattapaglia in melon (Cucumis melo L.). BMC Plant Biology 2011, 11:111. D, Vaillancourt R, Kilian A: A high-density Diversity Arrays Technology 58. Collard B, Mace E, McPhail M, Wenzl P, Cakir M, Fox G, Poulsen D, Jordan D: (DArT) microarray for genome-wide genotyping in Eucalyptus. Plant How accurate are the marker orders in crop linkage maps generated Methods 2010, 6:16. from large marker datasets? Crop and Pasture Science 2009, 60:362–372. Hudson et al. BMC Genomics 2012, 13:240 Page 11 of 11 http://www.biomedcentral.com/1471-2164/13/240

59. Grattapaglia D, Silva-Junior O, Kirst M, de Lima B, Faria D, Pappas G: High-throughput SNP genotyping in the highly heterozygous genome of Eucalyptus: assay success, polymorphism and transferability across species. BMC Plant Biology 2011, 11:65. 60. Faria D, Mamani E, Pappas G, Grattapaglia D: Genotyping systems for Eucalyptus based on tetra-, penta-, and hexanucleotide repeat EST microsatellites and their use for individual fingerprinting and assignment tests. Tree Genetics & Genomes 2011, 7:63–77. 61. Acuña C, Fernandez P, Villalba P, García M, Hopp H, Marcucci Poltri S: Discovery, validation, and in silico functional characterization of EST-SSR markers in Eucalyptus globulus. Tree Genetics & Genomes 2012, 8:289–301. 62. Sansaloni CP, Petroli CD, Jaccoud D, Carling J, Detering F, Grattapaglia D, Kilian A: Diversity Arrays Technology (DArT) and next-generation sequencing combined: genome-wide, high throughput, highly informative genotyping for molecular breeding of Eucalyptus. BMC Proceedings 2011, 5(Suppl 7):P54. 63. Arcade A, Labourdette A, Falque M, Mangin B, Chardon F, Charcosset A, Joets J: BioMercator: integrating genetic maps and QTL towards discovery of candidate genes. Bioinformatics 2004, 20:2324–2326. 64. de Givry S, Bouchez M, Chabrier P, Milan D, Schiex T: CarthaGene: multipopulation integrated genetic and radiation hybrid mapping. Bioinformatics 2005, 21:1703–1704. 65. MergeMap. http://alumni.cs.ucr.edu/~yonghui/mgmap.html.

doi:10.1186/1471-2164-13-240 Cite this article as: Hudson et al.: A reference linkage map for Eucalyptus. BMC Genomics 2012 13:240.

Submit your next manuscript to BioMed Central and take full advantage of:

• Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit